Duos Technologies Group, Inc. (DUOT) Earnings Call Transcript
May 18, 2026
Earnings Call Speaker Segments
Hello, everyone, and thanks for being with us on this call. We released our earnings press release earlier today, and late last week we submitted our 10-Q for Q1 2026. You can find copies in the investor relations area of our site. I'd urge everyone listening to review both the press releases and that 10-Q filing so you have a clearer picture of certain specifics we plan to cover on this morning's call. Broadly speaking, our Q1 results show us carrying on with the strategic shift toward a platform centered on data centers, with our Duos Edge AI Technology Solutions division now standing out as the main engine of growth. Results for the quarter, as anticipated, mirrored our ongoing move away from the legacy rail operation along with the deliberate wind down of the New APR Energy Asset Management Agreement, and that agreement was what chiefly drove revenue during the period. At the same time, we're still positioned to beat our $50 million revenue goal for this year, backed by our strategic partnership with Hydra Host and an expanding pipeline of AI infrastructure deployments. Before I dive into the exciting updates on our Duos Edge AI and Technology Solutions divisions, I want to first bring you up to speed on our rail technology and Duos Energy subsidiaries. Since our previous call, we've kept making headway on divesting the rail division. A fairness opinion regarding the rail division's value is currently being worked through by the company, and that process should carry into the second quarter. As we've said before, this was a considered choice that lets us redeploy capital, cut SG&A, and concentrate on opportunities with higher growth. We'll share more specifics as things progress. Now, on to Duos Energy Corporation, where we experienced a ramp down this quarter as reliance on Duos services diminished. To recap, in December 2024, Duos struck an Asset Management Agreement with New APR Energy aimed at locating new contracts to engineer, procure, construct, and operate fast power plants. That was key for us in enabling the data center business transition now in progress. As we've mentioned previously, the AMA wraps up later this year, though we'll keep a 5% equity stake in the parent of APR Energy. During Q1, the company posted $1.55 million in revenue alongside a cost of goods sold of roughly $544,000. That marked a step down from the prior period, and we anticipate it will keep winding down over the coming quarters. Now I'd like to talk about our data center strategy and our newer line of business, Duos Technology Solutions. When you're building and deploying data centers at scale, keeping costs in check and optimizing procurement matters greatly, given how capital-intense this market is. Being a smaller buyer compared with hyperscalers and large colocation companies, we required a more efficient approach to procuring equipment, and that's what led to the creation of the Duos Technology Solutions. With this division, we can lower what we spend on procurement for our very own installations, and at the same time build a fresh, asset-light income source that caters to enterprise clients, hyperscalers, and contractors. It's a pleasure to share that throughout the first quarter, Duos Technology Solutions gained real momentum. We closed deals with eight new sizable data center operators and grew our backlog to about $14 million, every bit of which should ship and be invoiced during 2026. As of today, our pipeline for the Technology Solutions stands several orders of magnitude above that backlog, which adds to our confidence in the outlook, particularly regarding the revenue ramp in the year's second half. This new business line has low overhead and scales well, and it is also backed by firm commitments from customers. We anticipate that revenue coming from the Technology Solutions will both substitute for the revenue from the new APR AMA and deliver improved margins. At this point I'd like to move our conversation to the heart of our newly data center-focused organization, Duos Edge AI. Need for edge computing and AI infrastructure keeps expanding quickly, and in our view Duos is positioned well to meet that need via our modular data center platform. After our recent capital raises, among them the $65 million in financing we closed in March, our balance sheet is considerably stronger, and we are well-capitalized to back near-term deployments along with future growth. Our priority during the first half of 2026 is to keep carrying out our sales strategy so we bring in new customers within our markets and fully utilize the capacity of each EDC. Over the quarter, we advanced considerably on two key revenue streams: GPU-as-a-Service and high-power colocation. Under the GPU-as-a-Service agreement we hold with Hydra Host, we plan to roll out 2,304 NVIDIA GPUs throughout our edge data center platform. This contract amounts to roughly $176 million in total revenue across a 36-month term, with total anticipated revenue of about $50 million, projected margins above 80%, and expected EBITDA of approximately $40 million. Notably, beyond the GPU-as-a-Service revenue, this partnership should produce incremental colocation revenue of about $25 million over the term, which further improves the overall economics of the relationship; we have already gotten a $15 million down payment, with a further $3 million deposit still pending. We are actively carrying out the initial deployments. We continue to expect revenue from this agreement to start ramping in the year's second half. In a separate matter, we won a high-power colocation contract to supply 4.8 MW of critical compute capacity in support of a leading hyperscaler's high-density GPU cluster. Taken together, these agreements mark a significant commercial inflection point, creating two complementary high-margin revenue streams and validating our edge data center platform at scale. Simultaneously, we are also witnessing rising demand for high-density data center capacity, pushed by AI and advanced compute workloads, with that demand now measured in megawatts instead of kilowatts. These higher power capacity EDCs ought to deliver considerably greater monthly recurring revenue for Duos. At present, Duos holds 10 MW under contract, with another 15 MW slated to come online during 2026. We keep growing our roster of edge data center prospects to meet rising demand tied to AI training, inference, and high-performance computing workloads. On the map, we are pushing into several parts of the country — Maryland, Iowa, Georgia, and Texas among them — as we shape the platform to serve enterprise clients alongside hyperscale ones. Inside the EDCs we already operate, we have started holding open houses for nearby residents and potential customers alike, giving them a chance to see firsthand how edge data centers deliver quicker connectivity, computing power close to home, and readiness for AI. A handful of these community efforts were announced not long ago, and we anticipate hosting a number more in the months ahead. Ever since we disclosed our latest contracts, inbound interest has been robust from hyperscalers, neocloud providers, and other customers needing compute at a large scale. To back a backlog and pipeline that keep growing, we are right now assessing fresh power partnerships capable of enabling green solutions and speeding up builds at our megawatt sites, and we look forward to sharing thrilling news on that front soon. To wrap up, we see Duos sitting at a critical turning point: we are shifting toward a business model with greater growth and wider margins, creating solid visibility via contracted opportunities and pipeline, and setting the company up to produce substantial revenue and EBITDA gains as 2026 unfolds. Now I want to hand things to our CFO, Leah Brown, to walk through our financials for the first quarter of 2026. Leah?
Thanks, Doug. For Duos, 2026 has gotten off to a start that's both encouraging and productive. Several landmark announcements came during the first quarter, along with strategic financing, strong growth in backlog, strategic investment, and meaningful movement toward becoming a stronger company that can scale more effectively. I'll take you through how we performed financially in the first quarter of 2026 and point out the key operational drivers behind those results. In Q1 2026, total consolidated revenue came in at roughly $2.7 million, versus $4.9 million during the first quarter of 2025. Looking at how total revenue for Q1 2026 was put together, it comes to around $44,000 from technology systems, $562,000 out of technology solutions, roughly $532,000 in services and consulting, $1.5 million tied to related party services and consulting agreements, plus about $30,000 of hosting revenue. The drop in revenue overall was driven primarily by the planned down draw under the asset management agreement, or AMA, with Duos Energy and New APR, which Doug mentioned before. Turning to gross margin: Q1 2026 brought much better results for the company, with gross profit of $1.6 million and a margin around 59%, which is a substantial gain compared with the prior-year period. What drove this was lower cost of goods sold, mostly tied to the AMA transition and the drop in costs connected to it. In both the first quarter of 2026 and the first quarter of 2025, the company also recorded roughly $900,000 of revenue stemming from its 5% non-voting equity stake in the ultimate parent of New APR. Since no cost of revenue is tied to that revenue, it came in at a 100% gross margin. During the first quarter of 2026, we posted a net loss of roughly $3.5 million, whereas the same quarter in 2025 showed a net loss of $2.1 million. The main reasons for that year-over-year widening were decreased revenues — stemming from a narrower scope of services that Duos Energy supplied under the AMA with New APR — along with increased operating expenses. As we've mentioned on earlier earnings calls, hitting positive adjusted EBITDA during Q3 and Q4 of last year marked significant milestones for the company, showing early gains from revenue scale and better margins; for Q1 2026, adjusted EBITDA came in at — $1.5 million. No adjusted EBITDA figure was reported for the prior year period. On a comparable basis, this mirrors the effect of the items we discussed previously. Even though adjusted EBITDA for the quarter came in negative, we anticipate better profitability once revenue picks up over the next several quarters. Turning now to our balance sheet: as of the close of Q1 2026, the company held $33 million in cash and cash equivalents. That figure rose sharply from where we stood on December 31st, 2025, thanks to the $65 million capital raise we completed in March, which bolstered our liquidity and improved our capacity to sustain operations while fully financing the investments we've planned under our arrangement with Hydra Host. By March 31st, 2026, Hydra Host had landed a customer, and in May 2026 that customer put down a $15 million deposit, with another $3 million still pending. Turning now to how we see 2026 shaping up: at the close of Q1, bookings were around $43.5 million in revenue, and we expect to recognize all of it before the year is out — that figure includes contracted backlog as well as awards we believe are close to being won. Beyond that, roughly $1.1 million of deferred revenue from contracted technology solutions, originally booked in 2025, will convert into revenue during 2026, providing a further boost to our results. That's all grounded in signed contracts and near-term orders that are already in motion and lined up to run across 2026. So we're reaffirming our call that total revenue for 2026 will come in above $50 million. Let me take a quick moment to show how we get from roughly $2.7 million of Q1 revenue up to that $50 million full-year number, since we know that's top of mind for our investors. The biggest piece is GPU-as-a-Service, where we're looking at about $26 million, with most of it landing in the back half as the project goes live and utilization climbs; separately, we expect around $26 million out of our technology solutions backlog, which gives us a firm committed base. That figure includes $2.9 million sitting in deferred revenue right now that gets recognized in the latter half of the year. We're also still on pace to convert $15 million of bookings into revenue during 2026, backed by another $25 million in backlog. Beyond that, we anticipate the rest of the guidance being recognized thanks to growing contributions from colocation and infrastructure services, fueled by customers expanding, fresh hosting deployments, ongoing capacity build-out, and new customer wins we're actively chasing. All these visible drivers together are what give us confidence we'll hit the full-year target. To say it again: because of how revenue recognition timing works, a big chunk of revenue should land in the second half of the year, and that's also when we expect to get back to positive adjusted EBITDA. Doug, over to you now for your closing remarks.
Thanks, Leah. During the opening quarter of 2026, we kept building on our momentum, carrying out our AI infrastructure plan and growing our edge data center presence. The accolades our company has earned over this year highlight how strong our position is and confirm that the direction we've chosen is the right one. In our view, the approach we've taken fits with a number of potent sector forces: the swift expansion of workloads powered by AI, rising need for infrastructure that is high density and energy efficient, movement toward secondary markets where power is available, plus a wider trend of adopting modular data center solutions that get deployed faster. Meanwhile, shifting necessities around power, cooling, and sustainability are each transforming the competitive field, which only strengthens how vital a solution that is scalable, cost-efficient, and quick to reach market has become. As we move through the rest of 2026, we have focus, discipline, and an expanding pipeline of opportunities, and we feel our position is strong enough to seize what the market offers. With that said, I'll turn things over for questions, everybody. Thanks. Operator?
Hello, I'm asking on behalf of Edward Woo. Given how things are going for you in the U.S. data center sector, are there any intentions to grow outside the country?
That's a fair question at this moment, since inquiries keep arriving from abroad, particularly South America. In fact, I was over in London just last week, and there's plenty of interest. At this point, what we're mainly concentrated on is continuing to demonstrate that the approach works here in the U.S.; and most likely we'll remain with delivering our 25 MW during this year alongside our 50 the following year within the U.S.
Thanks. I have one additional item. Given how robust U.S. demand is, are new players coming into this space, or has anything shifted regarding the competitive landscape?
To be honest, some things have been happening. Armada clearly operates in this area, though their strategy differs. Their model leans more toward privatization alongside Microsoft, yet the demand for inference is beginning to show up. You can also see that computing needs to happen closer to the user, and naturally power remains a problem. A great deal of activity is emerging around the modular route and targeting that 5 MW-10 MW band, which allows faster deployment.
Great. Thank you.
Hi. Good morning, guys. I appreciate the time. Doug, regarding the Hydra Host GPU-as-a-Service agreement, could you give us an update on where hardware deployment and site readiness stand? What I'm trying to figure out is whether revenue starts showing up in the third quarter or, you know, at some point later in the year.
Yeah, definitely. By Thursday last week, both NVIDIA and Supermicro had gotten all of it; Supermicro is handling the rack and stack, so those cabinets are going to be completely filled and sent out to the site. In fact, that cuts roughly three weeks from how long we typically wait. Hopefully, we're expecting construction to begin July 1st rather than August. All signs suggest that's where things are headed, which would put us a month ahead of plan.
That sounds excellent. I'd like to know something.
To begin with, that would come to $4.4 million in revenue. My apologies, please continue.
Great. Great. Do we potentially see this partnership expand to other locations?
The Hydra Host partnership?
Ye-yes.
Right. As we move ahead with the Hydra Host partnership, what we're seeing is that on this initial model, we rolled out GPU-as-a-Service, correct? The GPU was actually purchased by us. Going forward, that isn't the approach we'll take, though they certainly have a large number of customers — in fact, more than 12 customers — who have interest in 5 MW or 10 MW, where the GPU has been bought by other parties and now needs deploying. Our collaboration with Hydra Host is going to continue expanding, and that expansion will happen on the co-location side.
That's great, thank you. I've got just one more question. You increased certain expenses over the quarter. Should we anticipate that carrying on for the rest of 2026, or does the cost base as it stands right now, you know, basically sustain the growth you're expecting through the close of the year?
When you mention infrastructure costs, is that about the $6.5 million for each megawatt?
No, what I meant was — sorry about that, Doug. I believe you absorbed certain sales and marketing expenses during the quarter, plus perhaps a small amount of D&A. What I'm getting at is whether further OpEx increases are necessary going forward to sustain the top line.
I'll cover this briefly. Clearly, everyone investing with us on today's call understands the rail operation is being spun off. What's proven difficult is getting those two entities apart. Many people initially see us as a rail company, then after looking more closely and seeing our activities, they become quite enthusiastic and pleased. Our approach involved deploying significant capital early in the year, and we'll continue doing so as we enter Q2, in order to truly create distance and divide the two operations. Considerable marketing spending has gone toward this. Naturally, we brought on Gateway, and they're performing wonderfully for us — already we can observe incoming calls being handled properly, with investors receiving the correct message and holding appropriate expectations about our efforts. This will taper off somewhere around the July, August period, since our progress is strong. I believe we'll reduce that. It was certainly something we needed to address.
Great. Well, I appreciate the added color. This is very helpful.
Thanks.
Hi there. How are you viewing the money you'll put into CapEx across the coming 12 months?
Regarding what we'll spend on CapEx during the coming 12 months, our plan is to bring another 5 MW site online, and that comes to about $6 million-$6.5 million. To reach our target, we expect to spend $30 million across the next two quarters. To remain on schedule, we'll likely spend an additional $30 million near the end of the year. Naturally, we'll buy additional product for next year so we're certain to reach our number for next year. Things are proceeding as planned. With the funding we currently have, we're sufficiently capitalized to reach our 25 MW number this year, and that's achievable using the money available to us.
Yeah. That's obviously a real edge over rivals. Looking at strategy, the agreement you have starting later in the year appears to run 3 years. Once that deal wraps up, how are you thinking about what happens with those GPUs? Is there a possibility they might be renewed?
Yes. We're actually considering 2 possibilities right now. First, the market indicates — and this can shift — but the market indicates those GPUs will carry a value of $50 million-$58 million, within that band, once the deal ends; we have 2 choices. First, we could simply sell off those GPUs and return to a pure colo operation, and at that point we'd no longer be in the GPU business. Alternatively, we can go back to that same customer, which we understand is typical. We return to that customer, and while we won't capture 100% of the revenue we earned during the initial term, we'd likely get somewhere between 40%-60% of that typical revenue. We'll evaluate both paths; it simply comes down to circumstances. Should we need the capital for growth, we'd likely sell off those GPUs. As you know, that capital can be redeployed into infrastructure.
That helps. You brought up as well that there are colo chances through the Hydra partnership. At the start, you were the one purchasing the GPUs, though down the road, customers who already own GPUs might come to you for deployment. When that happens, what kind of duties would fall on you?
Right. So if we put up a 5 MW site — a modular one, 5 MW, let's say located in Iowa — then our job involves delivering power, cooling, and connectivity. Essentially we function as a colo, the same way a QTS or an Equinix does. Anyone — the large brick-and-mortars — we're simply quite tiny, and we handle every service; their own equipment comes from them; rack and stack comes from them; when it comes to compute, the infrastructure comes from them. On our side, we deliver infrastructure covering power, cooling, plus the reliability of five nines, the backup power, the generators; that's what our core business is.
All right. Regarding those possibilities, my guess is you'd be committing to longer-term lease agreements alongside prospective clients. Would that be the right way to look at it?
That's right. Usually, the durations run between 5 and 10 years. Clearly, we're much fonder of those kinds of arrangements.
Alright, that's great. Honestly, what you're doing here is impressive. Maybe just one more question, since I'm somewhat new to this story. Regarding Hydra Host — could you walk me through a bit of what they bring to the table?
Of course.
What exactly do they bring to the table? Thanks.
Yeah, for sure. Hydra Host is essentially a company providing GPU-as-a-Service. The GPUs themselves aren't theirs. Selling and supporting those GPUs is what they're really good at. In practice — say the hyperscalers are their client — they'll approach businesses such as mine, or investors, or operators of data centers who are looking for income from GPUs; those parties will purchase the hardware. Whoever buys it holds ownership. Hydra Host merely handles the GPU. Installation is done by them, and they oversee sales. In essence, they generate revenue for you while also providing GPU support. You're the one purchasing all the GPUs and absorbing that cost, yet what you get back is the income, structured as a revenue share. A modest slice of that revenue goes to them.
All right.
Essentially, this serves those who aren’t part of the GPU sector but wish to participate in GPU revenue.
That tracks. Regarding the duties you mentioned involving colo, power, interconnection, and so on — could you also clarify who you teamed up with for those, and what kind of experience they bring?
Absolutely. Our gear has support behind it. Clearly, that's Schneider Electric; a great deal of Schneider Electric equipment is what we deploy. Vertiv is another one we use; internally, our own crew, we've got about 22 people, and our job is monitoring through our NOCs. Two NOCs exist for us. Jacksonville has one, and Amarillo, Texas has the other. Around the clock, those NOCs keep an eye on the pods, 24 hours per day. That covers break fix. That means, say an AC unit fails; within 2 hours we dispatch. Anything constructed alongside our pods mirrors a Tier 3 data center. The term for it is N+1. A redundancy factor applies to every component, which gives you time for repairs. Should something fail, the business isn't harmed. The five nines are still being delivered by you. Repairs can be done in the time available. Dual generators are why we have them. An A and a B feed is what everything you see is called, and we keep all of that up; obviously, we do hand it off to contractors located in those markets, yet the dispatch is controlled by us, the contracts are controlled by us, all the servicing is controlled by us.
All right. As you move ahead and consider all the possibilities, or expand, this is what I've got written down. How are you thinking strategically about, you know, locating power opportunities?
In our search for power, we operate differently, correct? We don't walk into a town expecting a 100 MW. We head to places where power exists but isn't being used — what we refer to as stranded. Say they put up a substation in a city in Iowa, right? They construct that substation at 20 MW since they figure that community will expand. So there's genuinely surplus power sitting there. Our crew heads out, locates spots with 5 MW-10 MW of stranded power, then we sign a contract and act fast. Often, at the two sites we currently have, those were essentially former bit mining locations, meaning the power was already in place; those operators assumed they'd consume plenty of power, but they never did. At that site, 10 MW is sitting available. What we do involves signing a lease with the landowner directly, and we take control of it for 20 years. The landowner earns income from the power too. It works in their favor to have someone like us, who will genuinely consume that 5 MW-10 MW, rather than a bit miner whose usage fluctuates. One month it's 1 MW, two months — you know, that load isn't steady. They want to profit from power, and our approach appeals to them.
Thank you so much.
Yes, sir.
I truly appreciate that.
Hey there, Doug. My question is really about the 10 MW and the colocation side of things. You've got 10 MW signed, and your guidance for this year is 25, which means another 15 will get done. For next year, I believe you mentioned 40. Could you maybe walk us through what demand looks like out there? Which kinds of customers are after this? If the same customers who'd want those 100 MW sites you brought up earlier exist, then why would they come to you for something in the 5 MW or 10 MW range? Perhaps you can help us understand how that demand connects.
For sure. That's a great question. Thank you, Justin. Let me step back. After we brought the first one online with Hydra Host, calls started pouring in for 5 MW-10 MW . Having spent 30 years in this industry, I found myself asking, "Why, why is it that suddenly someone wants five or ten meg when everyone else out there is chasing gigawatt and 100 MW?" You know, the amount of power they're after is just insane. The reality is that the 5 MW-10 MW, even up to 20 MW, space is set to become the hot new space. It serves training as well as inference models. What's going on is they've got to get their GPU deployed, and get it deployed fast. Everyone is aware that people are sitting on GPU; they purchased them, they put money in. They've got to burn them. Getting 5 MW- 10 MW stood up fast is something we're capable of, within under six months. That's not achievable under the other models people are deploying. The Microsofts, the Googles of the world, every one of them, they're all moving into inference, and they understand they must grab these pockets to handle their inferencing. In the 5 MW- 10 MW range, I can tell you straight away, looking at my wall right here, I've got 21 neocloud. Were I to have 5 MW or 10 MW, they'd snap it up. That need exists; right now they're working to expand their networks for inferencing. At last, it's here. This has been discussed for some time. Now they're actually doing it. You can look at press releases from Google, showing how they're aiming at 20 sites at present, same story; the moment has come, and deployment can happen faster. Clearly, speed matters most at this point.
Understood. Excellent. I just have one more question regarding the balance sheet. Regarding that $33 million in cash, I believe Leah brought up the $15 million prepayment that was received, plus another $3 million coming. That happened in May, so is that extra cash on top of the $33 million you reported as of March 31st?
That's right.
Right, and only one more as well. I have a question. A filing appeared on the FTC site during the weekend showing that Elon Musk bought APR Energy. First, can you just verify whether that's the same APR Energy in which you hold a 5% stake? It would be great if you could give us any specifics there. Thank you.
Yeah. I'm not free to comment on that right now, but it is the same, clearly the same business. I can't talk about it at this time. With luck, we'll get some updates from them before long.
Understood. Thank you.
Hey, Doug — is my audio coming through now?
Yes, sir. How are you doing?
I'm fine, thanks. Yeah, that closing question was the very first one I noticed. Nothing has come out about it yet — no specifics at all.
Right.
From what I can tell, that's going to work in your favor financially. No matter how the specifics shake out, should this deal close, then your 5% cut of whatever that figure ends up being will flow into Duos, right?
That's right.
In the event that the sale goes through, will there be taxes on it? Is the figure we simply take 5% of — whatever that figure turns out to be — the figure that ends up appearing on your balance sheet? Do you have any sense of that? Is there anything you can carry forward? How would that appear?
Looking only at the money we'd get, there's a waterfall structure in the agreement. It isn't a simple matter of just taking the 5%.
Okay
on the transaction.
Yeah.
Sure.
Oh
Where taxes are concerned, I won't give you the exact figure, but let's just say..
Correct
he buys it for $100 million.
Right.
I realize there's no way for us to know precisely what that 5% amounts to. As for the 5%, will it be subject to taxation in the manner of a long-term capital gain, so that you end up netting out some portion of it? What I'm asking is simply that we speculate about the figure. Clearly, we have no certainty on it.
Yeah.
What I'm asking is, assuming it gets approved, how would that tax be structured?
Would it be a gross— Doug Recker: Yes.
Or is it a net figure? That's the question I'm putting to you.
At a broad level, I'd say we recognize that this constitutes a capital gain. We're not looking to disclose anything further.
Yeah
you know
Yeah
calculation around that transaction at this time.
Nico, I believe I can give you a somewhat clearer response to what you asked. We, well, you know.
Yeah
So as far as what we're doing with this shift in the rail business goes, I think..
Yeah
there are quite a few NOLs on hand. My guess is we're going to be fine, though the moment we have clarity, you'll hear from us.
My apologies — a sizable portion of what exactly? I simply missed that part.
It relates to how the rail shifted.
Yeah.
Yeah, yeah.
I simply couldn't catch your comment. A considerable quantity of what exactly?
Right, net operating losses. What I'm saying is, our company...
Oh, NOL, okay. Understood.
In that division, we have lost a great deal of money.
Great
we'll, uh-
No, I get it. That has to be the reason.
We ought to be able to confirm that. Yes.
So, from what's happening out of public view, you must be pretty thrilled, yeah? Because this is transforming the organization into the one you actually want going forward, isn't that right? Sort of centered on what you'd like to be doing when it comes to the data center.
You've got that exactly right. When you take a look at what we do, clearly, I stepped into this position, and I delivered this product into this company on behalf of our shareholders; nothing out there in the market beats it at the moment; all we have to do is split things apart and stay focused. Take a clear illustration, we, as you know, everyone listening to this call is aware of it. It's been difficult, agreed? That division consumed $900,000 of our money. We've got to deal with that, and that's what we're doing right here. I want to make you a promise that I'll get that wrapped up at the earliest opportunity; we're basically at the end of it now, so that has us feeling great; there's a lot of things this week that have us feeling great. Our level of excitement about the direction of this company is extremely high. This is just, you know, something like this is just, you know, one more quiver, one more arrow for our quiver. This is great stuff.
That brings me to what I wanted to ask next. You know, normally you're extremely energized during these calls, and this one feels somewhat more subdued, and that's even with the $2.7 million in revenue for the quarter. Clearly, a figure like that isn't what I believe anyone wants when they buy into this company. I'm wondering. Do you think this is — like, given that this pivot is taking place, the effort you're putting in, maybe you're recording it now, but the revenue hasn't been recognized yet. Is this, like, the pivot quarter or quarters? Like, is this something we should be anticipating or should have anticipated?
Yeah
You were expecting?
Yeah
So, all of that has been secured by you. This tech solutions work is sitting in the queue.
Correct.
since we're booking that amount this quarter, nobody ought to view $2 million and think, wow, that quarter was a total bust. The work you're putting in will eventually bring in money.
Yeah.
Over the coming few quarters and going ahead, in what way, if at all, does this figure not really matter, or is it something we ought to, say, raise questions about for this quarter — that's what I'm asking?
Right, that's a perfect illustration — had I been able to get all of those goods out the door, since with that kind of business you've got to deliver it, correct, in order for the revenue to be recognized.
Yes.
This quarter, I had the option to record $14 million, correct?
All right.
During the previous earnings call, I stated plainly that this revenue would begin to take off toward the end of Q2, Q3, and Q4; that has consistently been how our model works. What our model looks like at present, clearly, is to continue onward, to keep performing the way we have been, and to bring those additional sites online. You know, once I'm satisfied with that Iowa site, that represents a further 5 MW at $2 million per megawatt, correct?
Yeah.
You genuinely begin to notice this taking off. What we're doing is precisely what I hoped the team would do — build these things, get them out there, stay focused, keep your head down. The infrastructure division, keep it running. I'm telling you, forgive me if I don't sound excited, we are working around the clock, 24/7, and it is great work. Sorry about that.
No, no.
We are extremely pumped. Yeah.
So, on that $2 million, can I just pose the question?
Oh, right. No.
Iowa came up in what you said. Which part of Iowa is that project located in?
Muscatine is the name. I never pronounce it correctly. Muscatine, Iowa — there it is. Expect a press release from us within a day or two. Illinois is just next door, so I still get that low latency down to Kemah, which matters. That site is what our hyperscale customer is after, and the Texas one too. The Texas site sits just outside Amarillo.
Okay.
We have now teamed up with a company that will be a fantastic ally for us going ahead; in fact, they are already an ally.
I'm guessing this isn't a big metropolitan region in Iowa.
Nope, nope. Right. You know, you're just 20 miles away from a major market area. Yep.
Certainly. Certainly.
That is the location where the power got stranded. That is the location where the power exists. Honestly, if I had 10 MW of infrastructure, you know, rolling off the line, I could electrify 10 MW in that spot right now. It is available.
Okay.
It's transmission down; it's beautiful.
All right, all right. I believe I've got a couple additional items to ask about.
Of course.
So, regarding this entire backlog situation and revenue — the reality that the work gets done at present yet isn't reflected in the numbers. After that, well, I noticed a few backlog-related items. Actually, during the earlier portion of the Q&A, you mentioned putting 30 MW into service. You see, there seems to be quite a bit of a communication gap about whether we're actually deploying anything, and how that converts into revenue. Just attempting, purely so that things become, well, clearer.
Yes
Also, within the release itself, it mentioned that following the quarter, a customer prepaid you $15 million, correct?
That's right.
like,
Yeah
I'm just working through the various, the various ways it's been phrased and figuring out which figures go where.
Right.
I'm keeping my fingers crossed here — regarding the backlog figures, it appeared they had been split out by booked backlog. For data center, the figure was $43.5 million. Tech solutions, from what I could see, came in at $14 million. Do those figures hold up? That seems accurate?
Indeed, that is completely correct.
So does that mean the $43.5 million along with the $14 million represent contracted business, and if so, over what period — or is no period specified?
Within this year. Correct.
Or-
This year.
That's this year.
This year. Yeah. Yeah.
Okay.
This year. Yeah.
Going forward, will you begin reporting that as something broken down, similar to a backlog, from both of those perspectives?
Yes. Yes.
Yeah, that would be extremely useful if you could. Should you do so, it truly assists in making the narrative clearer.
Of course. Yeah.
So then-
Yeah
Regarding that $43.5 million of backlog, does it count the $15 million prepayment?
On the $15 million prepayment — and there is another $3 million currently pending — recognition will happen across the contract's duration, meaning the customer contract of three years. Booking $18 million right away isn't what will occur; instead it will be spread across the life of that customer contract.
What I'd like to ask about is that $15 million prepayment you called out in the release — does that represent cash flow coming in, and is it?
Yes
When you say that’s included, does it differ from the booked backlog you reported?
Our reported backlog does not include that.
Okay.
That won't get picked up until over the term of the agreement, which runs 3 years.
Yeah. You follow me, correct? It's only about attempting to understand things and obtain the sharpest view, and you pointed out that $15 million prepayment during the call, and after that I'm simply uncertain how that aligned with a few of those other figures.
Of course.
Then, you know.
Right
Regarding the guidance — as Doug mentioned, it concerns megawatts. You have a megawatt target, plus a revenue target too. My aim is simply to confirm we're consistently referring to identical metrics, rather than mixing things up. We're not discussing megawatts that are booked and built as opposed to revenue, since that, in my view, resembles a pendulum that keeps swinging one way and then the other throughout the discussion.
Correct.
so that everybody's aligned and your stock can be valued in light of all these positive steps you're taking. Just to make things as transparent as we can. From what I'm hearing, we may be right at the point where this rail car operation along with the gas-powered turbine business gets taken out, and my sense is that can only be beneficial, you know, to clear things up, so.
Without a doubt. Without a doubt.
Alright. The final item concerns the unit itself. You brought up competition — increased competition. On this call, Armada came up. You noted significant interest exists in that 5 MW-10 MW bracket. From my perspective, that's somewhat new information, simply because you folks are basically the only company actually doing this. That clean room represents an enormous competitive edge. I just want clarity on — I saw a video featuring them with a semi-mobile data center, and somebody is lowering one down onto a Navy ship. Is that essentially a PR move?
Yes. Yeah.
Is it real?
It's-
So, so-
No, it. Yeah.
Who?
Nico, it's real.
Which one of you is actually here? I mean, who's actually real?
That's a completely separate use case. Right.
Absolutely, absolutely.
This is an entirely separate use case from what we're working on. Right. When I mention folks entering the market, I was referring more to, you know, modular, correct?
Of course.
Deploying 3 MW, 4 MW, or 5 MW isn't something they're going to do. That's just not happening, and it's not aimed at serving multiple customers either, correct?
All right.
Those are solely for a single privatized customer. We handle quite a bit of work for the government. That was merely something I brought up as an illustration, yet you witness plenty these days, should you step into the industry and be part of it as we are.
All right.
A lot of 3D renderings are going to come your way. Plenty of folks will be out there claiming, "Hey, we're running inference. We're doing this." That explains why the 2 main hyperscalers showed up at our site — last week they took a tour of Corpus Christi — and the reason they traveled down was to see the thing with their own eyes. Keep in mind, though, that pod isn't the high-density one.
Sure.
They were looking for actual hands-on work to have been completed. They were interested in seeing the way we go about establishing the standard of that work. Both of them gave their approval.
Alright.
The way we capture the market is through exactly this approach. This has been my focus for 9 to 10 years at this point. Throughout my career, I've placed more than 30 of these into service. Take a look at the very first unit. Or check out the one we installed just 3 weeks back. We absolutely understand our craft, and while it isn't rocket science, we've mastered it, correct? We've truly mastered it.
What you said leaned more toward the upbeat side — that interest in this modular kind of concept has grown quite a bit, instead of.
Right
there's a lot more, like, competition coming.
Yeah.
All right.
I have confidence. Why do I feel that way, Nico? It comes down to this. First, it generates buzz for the sector, correct? Everyone is beginning to pay attention to it, and folks are beginning to take action. Second, I would compete against any one of them any day, and if you were in my sales team, you I have a single gimmick. This is my gimmick. When we approach a prospect to make a sale, like this is how we landed our initial hyper, I told them, "Listen, I will cover the cost for you to visit their pod. I will fly you out. We will tour it as a group, and afterward we will check out mine. If there is no pod available to view, then you must sign with me." I have not been defeated yet.
That's exactly what I was getting at with "fired up." This is precisely the thing I've been referring to.
Oh, my apologies.
No. You follow what I'm saying? I mean, I will take on anyone at all who claims that you have faith in your—.
That's right.
That's your unit. When I said who's real, it's basically what you just mentioned. Prototypes are everywhere, but the real question is who can truly make who can actually produce this at scale?
Yeah
So is it just you and Armada? For this particular space, are those really the only two you'd be up against? How do you and Armada differ? Is it the clean room? I mean, does that do it?
Yeah
to protect you if this. You know, if there's not really barriers.
Yeah
when it comes to producing a rectangular box, there's so much interest out there. Does that clean room give you adequate protection?
Yes, the clean room does provide sufficient protection, however what they're actually utilizing is closer to a shipping container, and 10 years ago I began my career working with those. High density isn't possible inside one of those.
All right.
The way they operate, it's a single customer with 1 to 2 cabinets. As for my customers, that's not something anyone is interested in.
All right.
Take Johnson & Johnson, a hospital, or a manufacturing plant — they simply require a compact unit for their own private compute and AI needs. That's their focus. Their market differs, yet they are rolling out modular. That's all. It isn't even a like-for-like comparison. I only wanted to point out that someone out there is doing modular.
Okay.
A lot of folks notice another party is engaged in that. That's a completely different situation, and it's more privatized. It does generate hype out there, correct? That's something I appreciate.
Correct. No, you've got a point. Search for it on Google, and you'll find this all over the place.
Yep.
You said you've completed 30 of these. You also said that if necessary, those GPUs can be sold once the contract wraps up, and I'm fairly certain that at some point during the past 18 months of conversations with you, you noted that the very first data center you constructed using one of these still holds all that older technology. It remains completely leased. Do you honestly believe?
Yep
Is it conceivable you'd ever part with the GPUs? I ask because, barring some strange turn of events in AI, there would surely be a buyer — that's what your remark about the very first unit you built 20 years ago suggests.
Abso-
or something along those lines.
For sure, there's going to be someone who wants it. Honestly, just to be completely transparent, at this moment there are folks interested — they're attempting to purchase our Hydra Host contract as we speak.
Oh, really?
Oh, yeah. Are you joking with me? We receive phone calls. Let me say this straight to you folks, if I possessed sufficient funds to roll out 10 MW sites, every one would be occupied. I'd list them at auction on a Tuesday. By Thursday, not a single one would remain.
Okay, so surely somebody listening to this gets it — a 10 MW unit throws off revenue big enough that it'd justify fronting, what, 3 years of that capital as a loan to you, whatever way the deal gets put together.
Exactly.
In any case, there's no need for you to respond to that. It was just a thought.
No, no.
Last one is.
Yeah.
Sure, feel free to start whenever you like.
Alright. Keep in mind what our approach has been. The plan we had was to pull this off and demonstrate our capability to do so — first of all — plus that we're able to house GPU, and that deployment can happen fast. Who were we trying to attract attention from? The folks at NVIDIA, Supermicro — those types. Honestly speaking, I have no desire to pursue additional equity. I really don't. If someone could provide backing for me, whether that's an NVIDIA or a Dell or a comparable company, who recognize the necessity — first off, they're recognizing it right now — and before long you'll witness them taking this step. I want to be at the front of the queue, since we're the initial ones out there raising a flag and saying, "Hey, take a look at what we've got; it's genuine. We're actually doing this. Feel free to speak with our customer; we possess $144 million in GPU value, so we're involved."
Well, if.
We
So when I take a look at you and ask myself, why exactly would you be having conversations with NVIDIA? Would it be fair to describe it this way — you'd be assisting NVIDIA since they're sitting on a backlog of GPUs that their customers already purchased but are unable to receive, given there's no place to install them, and you're able to complete these projects within 120 days, whereas these builds are getting delayed out in rural locations, the large data centers? If anything, they have an incentive to assist you in getting these projects constructed, because doing so actually frees up that backlog for them to recognize.
Exactly. You must have been slipping into my office and going through my playbook. Yes, that's correct. That's precisely it.
No, it's just that, well, logically speaking, it adds up. Those large data centers—
Naturally.
are getting pushed back.
Yeah.
In any case, you folks are able to complete those faster.
Correct.
to a company such as NVIDIA, or a similar one, you know.
Yes
Is the reality.
Yeah.
The final thing I want to ask about is power. Eventually, given everything that's getting constructed, we're going to run into the limit of what our electrical grid can handle. I'm certain your knowledge on this exceeds mine. At least, that's my hope. What, like, if this really is going to keep going at the speed that they claim it will for power, then something has to give. Like, I've come across plenty of proposals out there.
Yeah.
One item people treat almost like the holy grail, yet plenty of doubt surrounds it, is nuclear.
Yes.
that SMR thing, or whatever name people use for it, functions essentially as a power source, right?
Correct. That's right.
You know, first of all, is that even something that could actually happen? And second, once we reach the limit of what the electrical grid can handle, what's your sense of what goes down? Then, you know, the third thing is, the way I understand this stuff, sure, it's wonderful energy and it's efficient and uses little power, but at the same time it's kind of like a bomb just sitting right beside you.
Yes
regarding your hardware, such as the most critical items you're looking to have preserved.
Yes.
Even if it works.
Yeah
When it comes down to actual use, would folks really adopt it?
Yes. So here's the thing we don't usually say out loud, okay? Small modular reactors are going to require quite a while, okay? I can't picture many towns signing off on a nuclear or mini plant being built nearby. My sense is that's still years away.
All right.
Should that come to pass many years down the road, it would be a solid move. That's too distant for our focus. Our focus is on right now. If you caught it and you were listening when I held my earnings call, I brought up a different green company, correct? You've observed the folks, like those at Bloom, who've pulled this off for Google — essentially purchasing their output for the following 2 years. Natural gas is what powered those. On our end, deployment would likely take 6 months, 6 to 7 months. They're able to stand up 10 MW in a spot where the grid isn't involved whatsoever, and it's green to some degree. It's beneficial for the environment. It isn't harmful to the environment. No water is consumed. Natural gas is the whole of it. Those are the sorts of partnerships and items that we're driving toward, pursuing those routes, and with any luck you'll be hearing something before long. That's the path forward, and I'm saying it to you straight this moment.
Yeah.
So it has no impact on the nearby area, and those can be put into service, plus natural gas is what it runs on.
Alright, so here's a thought for your follow-up once the clean room is done — you ought to get a patent on this structure, with an electrical windmill mounted on top of the unit.
Yeah, understood.
Solar, solar panels.
Yeah, yeah.
And then, uh-
Well-
What I'm saying is, take the resistance to nuclear, and suppose it actually succeeded.
Right.
I think it's just new.
But-
What I was trying to get at is essentially public perception.
Right. Correct.
where like-
You got it.
When someone hears the word boiler, their immediate reaction is to assume it's a safe bet.
And that's right.
Boilers, you know, are constantly exploding.
That's right.
In any case, thanks— Well, sure. We can discuss that separately.
for the questions.
I'll, Yeah, I'll call you.
That's great, thanks.
All right. Appreciate it, Nico. Many thanks to you, sir.
Yeah.
Thanks. That previous discussion really helped a lot. The majority of what I wanted to ask has already been addressed, though I still have a couple left — one brief, one lengthy.
All right.
You mentioned that if the client holds ownership of the NVIDIA chips, then maintenance, power, and connectivity fall to you. Am I right in thinking that means your company bears the financial burden of linking these new facilities via fiber?
Yes, yes. The fiber carriers show up. Once a site is identified, we verify there's plenty of fiber in the surrounding area. Usually it's long-haul fiber, and we're beside a highway since that's where the fiber lines are laid, and then the carrier is the one that actually brings it in. Because as soon as the customer confirms they're located there, the carriers arrive since such huge bandwidth gets used; building into the infrastructure is worthwhile for them, plus they require multiple paths; the fiber itself isn't ours. Ownership of the connectivity isn't ours either. It belongs to the carrier, and they sell straight to the customer; that's the reason they come. After power is built and the customer puts pen to paper, that's the moment they arrive.
All right. The fiber isn't something you're accountable for, yet clearly you're putting these sites strategically in spots where the carrier's expense under the contract stays as low as possible. Right?
Precisely. That's absolutely right.
Thanks. All right, that clears things up. Incidentally, I traveled to Corpus Christi for the announcement, and being there truly made the situation click for me. You have a real gem in Bill Radford. He's terrific.
Oh, so you really did visit the site?
I did. I did.
Oh, excellent. Thank you.
He really helped me get a handle on what on earth is going on with this.
What did you think? Plenty of the folks who've put money in with us never get the chance to lay eyes on the pods themselves. How did it come across to you?
Two items really stood out to me as eye-opening. First, water isn't part of your process at all.
You've got that right.
This is a system that cools purely with air, though given how Bill arranges the stacks—he was walking through this—air is able to accomplish what it normally cannot in other centers, thanks to the manner in which he's set up the airflow; that made a difference.
Exactly. We contain the cold aisle and the hot aisle.
Sure. That clarified for me how you manage to do that while still delivering 100% availability. I got why the marginal costs come in so low. The center basically operates on its own. All you have to do is respond when issues come up.
Yep.
All right, that was really useful for me. I'm struggling to put all of this into a model, and maybe a call after this would work better. Let me just put that out there.
Sure. Give me a ring whenever. That way we can go through it without rushing. That would be excellent. We would be glad to have that conversation.
All right, excellent. I'll just lay out 3 things I'm assuming. You let me know whether I'm completely wrong about them. Does that work?
Absolutely.
Putting up a single one of those 10 MW sites costs somewhere in the neighborhood of $1 million to $1.4 million, correct?
That's wrong. Not even close. When you break it down per megawatt, the figure comes to about $6.5 million per megawatt. That $1.4 billion number belongs to the 300 kW units. I want to make sure that's understood.
Okay.
Yeah.
It's $0.5 million per MW. Okay.
Yeah. $6 million MW. Yeah.
So you're aiming to lease these out, and doing that within a sensible timeframe really doesn't look all that hard. When you lease it, the rate lands somewhere around 30%-50% of what it cost to build. Would you say that's roughly accurate?
It comes to $2 million for each MW.
Yeah. It's a little under $2 million per megawatt in revenue.
All right. All right. All right. You're spending $6.5 million for each megawatt, and you're bringing in $2 million for each megawatt. All right. That machinery appeared quite robust. I mean, roughly speaking, what kind of lifespan do you figure those units have?
Oh, the, the-
So there's a box they received, and it indicates roughly 10 years, correct?
Those are fine. When it comes to your generators, all of that is fine for 20 years. The sole component you'll be replacing is your batteries, correct? With your UPS, the batteries in those have a lifespan of eight to 10 years. Once you replace them, you're facing another injection of, I'd estimate on the high end, $80,000.
Okay, that really fires me up. Now, about the revenue numbers you're sharing — is that cash customers paid you to build out their centers, or is it coming from monthly lease payments? Is it all mixed together, and would it be clearer if I went through each line one by one?
No, no — it's what they pay us to lease. In the past, we'd charge a customer a single upfront installation fee to get them into the space; with our business model, the revenue that comes in is recurring; that's what a customer pays us each month to stay in that building. The payment is linked to power. How much power they use is what lets you figure out their revenue — 110% of it.
That revenue figure you're aiming for by the time the year wraps up — how much of it is going to recur annually? Are we talking half? 80%? I mean, after all, it's a lease.
Correct. The recurring
that's cash arriving on a monthly basis, correct?
that's right.
Yeah.
Yeah. Go ahead.
Right. That recurring revenue will last as long as the contract runs; in the case of our GPU-as-a-service offering, it’s a 3-year contract. When it comes to most of our co-location agreements, they range from 5 to 7.
Yeah
years. That's recurring.
Right. The sole items that don't repeat are, clearly, those on the infrastructure division side. That's the equipment portion, where we offer hardware to data center operators such as ourselves. You'll notice that.
All right. It's great to hear you don't want to rely on equity to fund this anymore, and I get that unexpected things come up. From your perspective, at what point will the free cash flow each month be enough to cover this expansion on its own? Is that 2 years away? 4 years away?
Our sector requires a tremendous amount of CapEx. Being in the data center space means capital expenditures are extremely heavy, correct? It all comes down to your particular approach. My approach involves continuous expansion. Here's the plan — and this is the reason behind our current actions — we will pursue debt financing; as an organization, we're on the verge of being able to do so. Once we bring another one of these facilities online, we ought to be capable of doing that without any trouble. Given our approach and the clients we're adding, that won't be a problem. That's the way we'll finance this company moving forward, since we will always be constructing new facilities. Because that's what you want to do. We resemble a REIT, don't we? Our goal is to hold as many assets as we possibly can, since that's where the worth lies. Perfect illustration: I put up that location in Iowa, and I put 10 MW there. After 4 years, 3 years, let's say those agreements run for 5 years. When 5 years pass, that client departs. I now have 10 MW open at that location. Every data center — every large brick-and-mortar data center — that I've possessed and divested over the years, none of them have declined, and all of them operate at full capacity. When you control the power and the infrastructure, it's incredibly valuable.
Alright, that puts my mind at ease quite a bit. As I head off, here's what I'm thinking — and yes, I'd like to accept what you proposed, just, you know, examining three or four separate forecasts.
That's exactly our line of work, sir. Feel free to ring us whenever.
the way I see it, once a year or two passes, you'll be bringing in sufficient cash flow that your margin reaches — just to pick a figure — no less than 15%, and you'll be funding things.
Oh.
With debt included, I'd hope that comes in below 10. Does that sound about right?
Right — if it were 15, we'd be in the wrong business. But yes.
All right, that's good news. In my mind, it ought to be 25 or higher, correct?
The rate we're borrowing at is 9%. Our margins ought to land in the 70% range.
All right, that's really useful, Doug, and I appreciate you putting that on the table.
Absolutely.
I'll follow up with you.
Yeah. We look forward to your call. Thank you, sir.
Hey there, morning everyone, and thanks for letting me ask. I believe you already addressed part of this, but could you give us some additional detail regarding how well the high power 1 to 2.5 EDCs have been doing? Could you explain a little further why buyers would choose those instead of the mega data centers at 100 MW+ that so many other companies keep bringing up? Like, which sorts of firms would be interested in them, and what are they ultimately used for?
Sure. First off, clearly, the massive sites you're seeing are those enormous training campuses, correct? The ones being constructed. Those facilities are gigantic and located far away, right? But when it comes to inference, you've got to be near the data, correct? These aren't for training, they're for inference, right? Computing has to happen where the, what we refer to as the eyeballs are. Wherever data is being generated and sent, you've got to be nearby; that's the reason you're beginning to observe the Googles of the world along with CoreWeave and all the rest standing up these inference locations. That has to be done relatively close to the data. Which is why you'll notice us moving into those Tier three markets. A perfect illustration, Corpus Christi. From Corpus Christi, I can still reach Houston in 5 milliseconds. Amarillo, same situation. Waco. That's still attainable in those locations, and those locations require inference as well. Community hospitals, local government, gaming—everything in that category—you're starting to notice that information needs to reside in proximity to the users. This has been observable for a long time, though now it is truly accelerating, with massive data quantities and processing power needing to be managed on-site. Which explains why, in earlier times, a single rack maxed out around 10 kW. With the computing capability we have now, it's 100 kW for each rack. An ordinary data center cannot accommodate that. You have to purpose-build for this tier. That's why there are so many new facilities being constructed specifically for inference work. It's coming — you'll witness a great deal of it. I'm keeping a very close eye on Google, since I'd really like the chance to sit down with them and say: here's what we bring — we can handle this on your behalf. You're already doing it regardless. Being in this business isn't what you want. We handle it, and we handle it skillfully. What you're after is an OpEx approach rather than CapEx; that's something we're able to provide. Their interest is purely in computation; running and upkeeping buildings isn't something they want on their plate. Naturally, this will expand and all of that — and clearly, when it comes to this area, you folks are among the very best. You all have a deeper understanding of this sector than nearly anyone else does, and you can tell which direction it's heading when it comes to inference and how much faster deployment becomes possible. Everywhere around the world, neoclouds are keen to put their GPUs to work generating returns, and this hands them a speedier route to get that done as well. I hope that's clear.
Yes. Thanks for that; it's useful, and I appreciate it. I suppose, I believe you've already addressed revenue per megawatt on colo, but when it comes to the economics of scaling, could you go over EBITDA margins once more, you know, given that CapEx and revenue? EBITDA margin per megawatt.
Sure.
Yes. For the entire year, our EBITDA margin came in at 17%. On an adjusted basis, EBITDA was 27% for the full year on a consolidated basis.
Great. As for the colocation side specifically, can you give us that as a separate figure?
Yes. When it comes to high-powered colocation, we see EBITDA margins of roughly 80%. GPU-as-a-Service lands at about the same level.
That's great. Okay, I really appreciate it.
I appreciate it.
Hi there, and thanks. The majority of what I wanted to ask has already been covered. Congrats on the quarter. To sort of put it together in my own mind, it seems like you're moving from 25 MW — 40 MW in the current year to the following year. You said that with capital available, you could sell a great deal more systems, and you noted that you'd go to the debt markets to raise it. Do you figure a strategic investor could also be found? And last, I believe part of why the stock confuses people is that this company gets misunderstood; there are so many moving pieces. You're shifting business models, divesting some businesses. When do you expect coverage to arrive for the stock, equity coverage? Thanks.
Thanks for asking. For that final question, quickly — I believe it'll be soon, since what's happening right now is that people are beginning to actually see our deployments. The revenue is beginning to show up for them. Clearly, when this Hydra Host deal lands, hopefully within the coming 30, 40 business days, that will truly be a driver. We've spoken with a number of analysts, and we're in the midst of working with them at this moment; they've come in and performed a deep dive. At this point they get it. Even those analysts had no understanding of what we do, and they're assisting us in getting the story out. My prediction is that, hopefully, within the coming 30 days we see something. Additional drivers are on the way that will compel them to come visit us. I'm eager for that too.
All right, understood. And then, regarding a strategic investor — that was my second question. Could that be a possibility?
Yes.
as opposed to needing to go out and secure additional capital?
For sure; that's the reason we're — I want to phrase it as we're still operating under a proof of concept approach, though honestly that isn't really accurate anymore, is it? We're in execution mode now; this is the moment, and getting in front of companies like NVIDIA — ones that could stand behind us on an initiative like this — has been my responsibility, since they appreciate what we're building; they recognize that inference clearly represents the future direction; they're looking to move their GPU inventory; plus they've got plenty of GPU clients who are holding onto capacity; they're eager to free that up so those clients can purchase additional units; that company would serve as one strategic partner of mine, and beyond them, the entities financing the major data centers, correct? Blackstone along with every other large player, including DigitalBridge. Each of them holds positions in these massive data centers, correct? Vantage as well as DataBank. What lies ahead for those data centers? They'll view us through a hub and spoke lens. Clearly they're running training modules, yet they'll be seeking inference modules deployed in the field. By hub and spoke, I mean you've got your primary data centers in Dallas, yet nothing exists in Corpus Christi. Nothing exists in Waco, in those suburbs, in those more distant, smaller Tier 3 markets, Tier 2 markets. Deploy these in those locations, and they'll purchase them because now they possess that hub and spoke structure enabling them to offer their services there, and that client, 90% of the time that client sitting in that pod also sits in their central data center. I foresee, much like with my initial company, EdgePresence, we accomplished this previously and DataBank provided our funding. DataBank gave us $35 million to validate that concept. Sadly, we got acquired before we could complete that validation. That's where I envision this partnership emerging, and it's entirely logical, and pursuing it that way would be entirely logical.
Fantastic. Thanks so much. I can't wait for that coverage, honestly — it'll help me get a clearer picture, and the street will benefit too.
Absolutely.
Thanks once more.
Same here. Same here, believe me. I can only handle a certain number of calls each day. I've still got selling to do.
Yep.
Thank you for your question.
What matters more is getting out there and making sales. Thanks.
Thanks.
Alright.
I'd like to express my gratitude to all of you, and hopefully what we've laid out gives you a solid picture of the direction we're headed — we're operating at 150%. You ought to feel proud of what we've accomplished. We'll continue pushing forward, and there's excitement on our end about some favorable announcements coming this week too. I'm eager to speak with every one of you. Feel free to ring me whenever you have questions. I'm available around the clock. I truly appreciate all your backing. We're looking forward to connecting with you before long. Many thanks to each of you.
Thank you. Prior to wrapping up the call today, I want to read out Duos' safe harbor statement, which contains significant caveats about forward-looking statements delivered on this call. Statements of a forward-looking nature fall within the meaning of the Private Securities Litigation Reform Act of 1995 and appear in this earnings call. Words such as believes, expects, may, will, should, anticipates, plans, along with their negatives or comparable phrases, serve to flag forward-looking statements. Please be advised that such statements do not promise future results or occurrences, and they face numerous uncertainties, risks, and other forces, a great many of which lie outside our control. Those factors could affect how accurate the statements and the projections underpinning them turn out to be, and they could lead Duos Technologies Group, Inc.'s actual results to diverge materially from what the forward-looking statements anticipate. Among these risks and uncertainties — though not limited to them — are the ones set out under Item 1A in Duos' annual report on Form 10-K, which is expressly incorporated herein by reference, plus additional factors that may be described from time to time in Duos' filings with the SEC. We appreciate your participation today in Duos Technologies Group's first quarter 2026 earnings call. You may now disconnect.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Duos Technologies Group, Inc. transcript - plus 131,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Duos Technologies Group, Inc. earnings transcripts and 131,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.